From owner-svn-src-stable@freebsd.org Mon Jul 18 08:20:33 2016 Return-Path: Delivered-To: svn-src-stable@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 35726B9C1E9; Mon, 18 Jul 2016 08:20:33 +0000 (UTC) (envelope-from jch@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id DD404141A; Mon, 18 Jul 2016 08:20:32 +0000 (UTC) (envelope-from jch@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id u6I8KWX0037799; Mon, 18 Jul 2016 08:20:32 GMT (envelope-from jch@FreeBSD.org) Received: (from jch@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id u6I8KVfd037797; Mon, 18 Jul 2016 08:20:31 GMT (envelope-from jch@FreeBSD.org) Message-Id: <201607180820.u6I8KVfd037797@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: jch set sender to jch@FreeBSD.org using -f From: Julien Charbon Date: Mon, 18 Jul 2016 08:20:31 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-10@freebsd.org Subject: svn commit: r302995 - stable/10/sys/netinet X-SVN-Group: stable-10 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-stable@freebsd.org X-Mailman-Version: 2.1.22 Precedence: list List-Id: SVN commit messages for all the -stable branches of the src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 18 Jul 2016 08:20:33 -0000 Author: jch Date: Mon Jul 18 08:20:31 2016 New Revision: 302995 URL: https://svnweb.freebsd.org/changeset/base/302995 Log: MFC r261242: Decrease lock contention within the TCP accept case by removing the INP_INFO lock from tcp_usr_accept. As the PR/patch states this was following the advice already in the code. See the PR below for a full discussion of this change and its measured effects. PR: 183659 Submitted by: Julien Charbon Reviewed by: jhb Modified: stable/10/sys/netinet/tcp_syncache.c stable/10/sys/netinet/tcp_usrreq.c Directory Properties: stable/10/ (props changed) Modified: stable/10/sys/netinet/tcp_syncache.c ============================================================================== --- stable/10/sys/netinet/tcp_syncache.c Mon Jul 18 07:06:44 2016 (r302994) +++ stable/10/sys/netinet/tcp_syncache.c Mon Jul 18 08:20:31 2016 (r302995) @@ -683,7 +683,7 @@ syncache_socket(struct syncache *sc, str * connection when the SYN arrived. If we can't create * the connection, abort it. */ - so = sonewconn(lso, SS_ISCONNECTED); + so = sonewconn(lso, 0); if (so == NULL) { /* * Drop the connection; we will either send a RST or @@ -922,6 +922,8 @@ syncache_socket(struct syncache *sc, str INP_WUNLOCK(inp); + soisconnected(so); + TCPSTAT_INC(tcps_accepts); return (so); Modified: stable/10/sys/netinet/tcp_usrreq.c ============================================================================== --- stable/10/sys/netinet/tcp_usrreq.c Mon Jul 18 07:06:44 2016 (r302994) +++ stable/10/sys/netinet/tcp_usrreq.c Mon Jul 18 08:20:31 2016 (r302995) @@ -643,13 +643,6 @@ out: /* * Accept a connection. Essentially all the work is done at higher levels; * just return the address of the peer, storing through addr. - * - * The rationale for acquiring the tcbinfo lock here is somewhat complicated, - * and is described in detail in the commit log entry for r175612. Acquiring - * it delays an accept(2) racing with sonewconn(), which inserts the socket - * before the inpcb address/port fields are initialized. A better fix would - * prevent the socket from being placed in the listen queue until all fields - * are fully initialized. */ static int tcp_usr_accept(struct socket *so, struct sockaddr **nam) @@ -666,7 +659,6 @@ tcp_usr_accept(struct socket *so, struct inp = sotoinpcb(so); KASSERT(inp != NULL, ("tcp_usr_accept: inp == NULL")); - INP_INFO_RLOCK(&V_tcbinfo); INP_WLOCK(inp); if (inp->inp_flags & (INP_TIMEWAIT | INP_DROPPED)) { error = ECONNABORTED; @@ -686,7 +678,6 @@ tcp_usr_accept(struct socket *so, struct out: TCPDEBUG2(PRU_ACCEPT); INP_WUNLOCK(inp); - INP_INFO_RUNLOCK(&V_tcbinfo); if (error == 0) *nam = in_sockaddr(port, &addr); return error;